Skip to content

Commit f24e7e6

Browse files
committed
Merge 3.2: issue #2892
2 parents f588dd6 + 91d5193 commit f24e7e6

4 files changed

Lines changed: 49 additions & 30 deletions

File tree

Lib/test/test_xml_etree.py

Lines changed: 1 addition & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -754,6 +754,7 @@ def iterparse():
754754
... print(action, elem.tag)
755755
... except ET.ParseError as v:
756756
... print(v)
757+
end document
757758
junk after document element: line 1, column 12
758759
"""
759760

Lib/xml/etree/ElementTree.py

Lines changed: 23 additions & 15 deletions
Original file line numberDiff line numberDiff line change
@@ -1250,6 +1250,7 @@ def __init__(self, source, events, parser, close_source=False):
12501250
self._close_file = close_source
12511251
self._events = []
12521252
self._index = 0
1253+
self._error = None
12531254
self.root = self._root = None
12541255
self._parser = parser
12551256
# wire up the parser for event reporting
@@ -1291,24 +1292,31 @@ def __next__(self):
12911292
while 1:
12921293
try:
12931294
item = self._events[self._index]
1295+
self._index += 1
1296+
return item
12941297
except IndexError:
1295-
if self._parser is None:
1296-
self.root = self._root
1297-
if self._close_file:
1298-
self._file.close()
1299-
raise StopIteration
1300-
# load event buffer
1301-
del self._events[:]
1302-
self._index = 0
1303-
data = self._file.read(16384)
1304-
if data:
1298+
pass
1299+
if self._error:
1300+
e = self._error
1301+
self._error = None
1302+
raise e
1303+
if self._parser is None:
1304+
self.root = self._root
1305+
if self._close_file:
1306+
self._file.close()
1307+
raise StopIteration
1308+
# load event buffer
1309+
del self._events[:]
1310+
self._index = 0
1311+
data = self._file.read(16384)
1312+
if data:
1313+
try:
13051314
self._parser.feed(data)
1306-
else:
1307-
self._root = self._parser.close()
1308-
self._parser = None
1315+
except SyntaxError as exc:
1316+
self._error = exc
13091317
else:
1310-
self._index = self._index + 1
1311-
return item
1318+
self._root = self._parser.close()
1319+
self._parser = None
13121320

13131321
def __iter__(self):
13141322
return self

Misc/NEWS

Lines changed: 2 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -350,6 +350,8 @@ Core and Builtins
350350
Library
351351
-------
352352

353+
- Issue #2892: preserve iterparse events in case of SyntaxError.
354+
353355
- Issue #13287: urllib.request and urllib.error now contains a __all__ and
354356
exposes only relevant Classes, Functions. Patch by Florent Xicluna.
355357

Modules/_elementtree.c

Lines changed: 23 additions & 15 deletions
Original file line numberDiff line numberDiff line change
@@ -3016,6 +3016,7 @@ PyInit__elementtree(void)
30163016
" self._file = file\n"
30173017
" self._events = []\n"
30183018
" self._index = 0\n"
3019+
" self._error = None\n"
30193020
" self.root = self._root = None\n"
30203021
" b = cElementTree.TreeBuilder()\n"
30213022
" self._parser = cElementTree.XMLParser(b)\n"
@@ -3024,24 +3025,31 @@ PyInit__elementtree(void)
30243025
" while 1:\n"
30253026
" try:\n"
30263027
" item = self._events[self._index]\n"
3028+
" self._index += 1\n"
3029+
" return item\n"
30273030
" except IndexError:\n"
3028-
" if self._parser is None:\n"
3029-
" self.root = self._root\n"
3030-
" if self._close_file:\n"
3031-
" self._file.close()\n"
3032-
" raise StopIteration\n"
3033-
" # load event buffer\n"
3034-
" del self._events[:]\n"
3035-
" self._index = 0\n"
3036-
" data = self._file.read(16384)\n"
3037-
" if data:\n"
3031+
" pass\n"
3032+
" if self._error:\n"
3033+
" e = self._error\n"
3034+
" self._error = None\n"
3035+
" raise e\n"
3036+
" if self._parser is None:\n"
3037+
" self.root = self._root\n"
3038+
" if self._close_file:\n"
3039+
" self._file.close()\n"
3040+
" raise StopIteration\n"
3041+
" # load event buffer\n"
3042+
" del self._events[:]\n"
3043+
" self._index = 0\n"
3044+
" data = self._file.read(16384)\n"
3045+
" if data:\n"
3046+
" try:\n"
30383047
" self._parser.feed(data)\n"
3039-
" else:\n"
3040-
" self._root = self._parser.close()\n"
3041-
" self._parser = None\n"
3048+
" except SyntaxError as exc:\n"
3049+
" self._error = exc\n"
30423050
" else:\n"
3043-
" self._index = self._index + 1\n"
3044-
" return item\n"
3051+
" self._root = self._parser.close()\n"
3052+
" self._parser = None\n"
30453053
" def __iter__(self):\n"
30463054
" return self\n"
30473055
"cElementTree.iterparse = iterparse\n"

0 commit comments

Comments
 (0)